set plane coordinates

[jburel]

This PR fixes an issue when setting the plane associated to a given shape.
The current implementation does not handle images with for example t=1, z=1 and c>1.
This is based on https://github.com/ome/omero-insight/blob/master/src/main/java/org/openmicroscopy/shoola/util/roi/io/ROIReader.java#L227

[ppouchin]

Ok. It fails because in some tests, ROIs exist without images.
Maybe the "position -> C / Z / T" conversion would work when saving the ROI to an image?

[jburel]

If the image is accessible from the roi it will work
roi.getImage() will not be null and it will allow to perform the check
Unfortunately the ImagePlus object is null in the current implementation.

[ppouchin]

I have a proposition (currently running test locally):

    protected void copyFromIJRoi(ij.gui.Roi ijRoi) {
        LengthI size          = new LengthI(ijRoi.getStrokeWidth(), UnitsLength.POINT);
        Color   defaultStroke = Optional.ofNullable(Roi.getColor()).orElse(Color.YELLOW);
        Color   defaultFill   = Optional.ofNullable(Roi.getDefaultFillColor()).orElse(TRANSPARENT);
        Color   stroke        = Optional.ofNullable(ijRoi.getStrokeColor()).orElse(defaultStroke);
        Color   fill          = Optional.ofNullable(ijRoi.getFillColor()).orElse(defaultFill);
        data.getShapeSettings().setStrokeWidth(size);
        data.getShapeSettings().setStroke(stroke);
        data.getShapeSettings().setFill(fill);
        
        // Set the plane 
        int c = ijRoi.getCPosition();
        int z = ijRoi.getZPosition();
        int t = ijRoi.getTPosition();
        
        ij.ImagePlus ip = ij.WindowManager.getImage(ijRoi.getImageID()); //ijRoi.getImage() returns null
        if (ip != null) {
            int pos       = ijRoi.getPosition();
            int stackSize = ip.getStackSize();
            int imageC    = ip.getNChannels();
            int imageT    = ip.getNFrames();
            int imageZ    = ip.getNSlices();
            if (stackSize == imageZ) {
                z = pos;
                //reset values
                c = 1;
                t = 1;
            } else if (stackSize == imageC) {
                c = pos;
                //reset values
                z = 1;
                t = 1;
            } else if (stackSize == imageT) {
                t = pos;
                //reset values
                c = 1;
                z = 1;
            }
        }
        data.setC(Math.max(-1, c - 1));
        data.setZ(Math.max(-1, z - 1));
        data.setT(Math.max(-1, t - 1));
    }

[jburel]

Feel free to close the PR if your suggestion works or adjust my commits

[ppouchin]

Ok. I need to check in ImageJ if it works to be sure, but the tests pass.

[codecov]

Codecov Report

Patch coverage: 96.00% and project coverage change: -0.04 ⚠️

Comparison is base (f831aa0) 95.53% compared to head (7ce4bfe) 95.50%.

Additional details and impacted files

@@             Coverage Diff              @@
##               main      #65      +/-   ##
============================================
- Coverage     95.53%   95.50%   -0.04%     
- Complexity     1139     1143       +4     
============================================
  Files            45       45              
  Lines          3026     3048      +22     
  Branches        236      242       +6     
============================================
+ Hits           2891     2911      +20     
  Misses           78       78              
- Partials         57       59       +2     

Impacted Files	Coverage Δ
...n/java/fr/igred/omero/roi/GenericShapeWrapper.java	97.36% <96.00%> (-0.33%)	⬇️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[jburel]

@ppouchin I will do some tests with the workflow I have in place already later on

[jburel]

With the latest set of commits, using the workflow I used for testing, the problem is no longer visible